Assessment of Brown Bear\'s (Ursus arctos syriacus) Winter Habitat Using Geographically Weighted Regression and Generalized Linear Model in South of Iran

Abstract

Winter dens are one of the important components of brown bear's (Ursus arctos syriacus) habitat, affecting their reproduction and survival. Therefore identification of factors affecting the habitat selection and suitable denning areas in the conservation of our largest carnivore is necessary. We used Geographically Weighted Logistic Regression (GWLR) and Generalized Linear Model (GLM) for modeling suitability of denning habitat in Kouhkhom region in Fars province. In the present research, 20 dens (presence locations) and 20 caves where signs of bear were not found (absence locations) were used as dependent variables and six environmental factors were used for each location as independent variables. The results of GLM showed that variables of distance to settlements, altitude, and distance to water were the most important parameters affecting suitability of the brown bear's denning habitat. The results of GWLR showed the significant local variations in the relationship between occurrence of brown bear dens and the variable of distance to settlements. Based on the results of both models, suitable habitats for denning of the species are impassable areas in the mountains and inaccessible for humans.

Keywords

• Brown bear (Ursus arctos)
• Fars province
• Generalized Linear Model (GLM)
• Geographically Weighted Regression (GWR)
• Winter habitat suitability.